Formerly About Advanced Micro Devices Message Board

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Politics : Formerly About Advanced Micro Devices

Public Reply Prvt Reply Mark as Last Read File

Previous 10 Next 10 Previous Next

To: Paul Engel who wrote (31716)	4/12/1998 10:42:00 PM
From: Maxwell	Read Replies (1) of 1573683

Dr. Engel: <<At the ISSCC in February, IBM described their experimental PowerPC chip built on a 0.20 process with 6 layers of copper - and it reached a peak speed of 480 MHz.>> The reason that chip was unable to reach beyond 480MHz was due to front end limited, not backend. Front end limited means that the response time per clock frequency is gate limited rather than backend cross-talking among the metal lines. If you go back and read IBM 1GHz chip, that chip was fabricated using conventional Al/tungsten via technology. However there was an architectural advancement made. IBM made the gates very short and execute in parallel. That architectural advancement is what made the chip goes beyond the 1GHz range. Copper technology currently doesn't improve the speed much without architectural improvement. However copper has 50% better conductivity than aluminum and better electromigration properties. 50% better conductivity means that the cross-section of the metal line can be reduced by 50% over Al and still maintaining the same conductivity. Shrinking in metal lines mean you can shrink the chip more. The other nice feature of Cu is that the process is nicer to work with. Cu is currently deposited as a batch process using electroplating at low temperature. Aluminum on the other hand is using PVD above 250C and is a single wafer process. Electroplating is cleaner than PVD and throughput is unbelievable. Etching metal is a nightmare when the spacing drops below .40um. Etching oxide gaps of less than 0.4um is cake walk. There is nothing easier than etching oxide. The only thing I can say now is that companies that get the copper process down will get the upper hand in terms of manufacturing costs. Performance will come later. Maxwell

Report TOU Violation

Share This Post

Public Reply Prvt Reply Mark as Last Read File

Previous 10 Next 10 Previous Next